Transient analysis of temporary overvoltage and cable faults in underground medium voltage systems
Ricardo Manuel Arias Velásquez
Abstract
• In PV solar plants identified peak overvoltage levels that can reach up to 3.5 p.u. under fault conditions. • Circulating currents can contribute to over 15 % of the total fault current, with potential impacts on cable insulation and system protection. • The relative error in underground cables increases from 0.2 % below 1.8 km to 5 % at 5.4 km. • Homopolar current revealed an improvement in fault detection precision by 20 %, with traditional method. • During a fault, the current reaches 4.1 kA, with a screen current of 4.56 A and a corresponding voltage of 3.5 kV. • A new screen cross-section demonstrates a reduction in thermal stress by over 30 %. This paper evaluates the challenges faced by renewable power plants in 2024 concerning underground cables designed according to industry standards. It emphasizes the need for optimization in cable operation to enhance efficiency and power throughput capacity, given the substantial investment costs involved. The study highlights the interest in single-conductor cables and the associated issues of induced voltages and currents. A key focus is on inrush current and its impact on transmission system protection, particularly relevant due to the behavioral differences between inverters and synchronous machines, and inrush currents in the transformers of PV solar plants. Additionally, it addresses the objectives of managing AC overvoltage and overcurrent during faults and the significance of transient modeling in these scenarios. Our findings reflect the issues of induced voltage, circulating current, and inrush current in long underground cables, particularly in the context of large PV solar plants with numerous transformers. We explore the effects on splices and terminals and discuss the use of insulated-gate bipolar transistors (IGBTs) in transformers, as well as the impact of cable configuration on circulating currents and ferroresonance. Furthermore, the paper presents findings from real fault scenarios, demonstrating the influence of cable core cross-section and the increase in error with cable length. It highlights the importance of managing induced voltage, circulating current, and ferroresonance in underground cables, particularly at splices and terminals. The study reports a relative error of 0.2 % for a cable length of 1.8 km, increasing to 5.0 % for a length of 5.4 km in the modeling.